sensors
Article
Blockchain-Based Smart Home Networks Security Empowered
with Fused Machine Learning
Muhammad Sajid Farooq 1,† , Safiullah Khan 2,† , Abdur Rehman 1 , Sagheer Abbas 1 , Muhammad Adnan Khan 3, *
and Seong Oun Hwang 4, *
Citation: Farooq, M.S.; Khan, S.;
Rehman, A.; Abbas, S.; Khan, M.A.;
Hwang, S.O. Blockchain-Based Smart
Home Networks Security
Empowered with Fused Machine
Learning. Sensors 2022, 22, 4522.
https://doi.org/10.3390/s22124522
Academic Editor: Wenjuan Li
Received: 3 June 2022
Accepted: 14 June 2022
Published: 15 June 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Sensors 2022, 22, 4522. https://doi.org/10.3390/s22124522
1 School of Computer Science, National College of Business Administration & Economics,
Lahore 54000, Pakistan; msajidfarooq@lgu.edu.pk (M.S.F.); arbhatti@ncbae.edu.pk (A.R.);
dr.sagheer@ncbae.edu.pk (S.A.)
2 Department of IT Convergence Engineering, Gachon University, Seongnam 13120, Korea; safi@gachon.ac.kr
3 Pattern Recognition and Machine Learning Lab, Department of Software, Gachon University,
Seongnam 13557, Korea
4 Department of Computer Engineering, Gachon University, Seongnam 13120, Korea
* Correspondence: adnan@gachon.ac.kr (M.A.K.); sohwang@gachon.ac.kr (S.O.H.)
† These authors contributed equally to this work.
Abstract: Security and privacy in the Internet of Things (IoT) other significant challenges, primarily
because of the vast scale and deployment of IoT networks. Blockchain-based solutions support
decentralized protection and privacy. In this study, a private blockchain-based smart home network
architecture for estimating intrusion detection empowered with a Fused Real-Time Sequential Deep
Extreme Learning Machine (RTS-DELM) system model is proposed. This study investigates the
methodology of RTS-DELM implemented in blockchain-based smart homes to detect any malicious
activity. The approach of data fusion and the decision level fusion technique are also implemented to
achieve enhanced accuracy. This study examines the numerous key components and features of the
smart home network framework more extensively. The Fused RTS-DELM technique achieves a very
significant level of stability with a low error rate for any intrusion activity in smart home networks.
The simulation findings indicate that this suggested technique successfully optimizes smart home
networks for monitoring and detecting harmful or intrusive activities.
Keywords: Real-Time Sequential Deep Extreme Learning Machine; data fusion; blockchain; smart
home
1. Introduction
A smart home is connected to the Internet, allowing users to manage a variety of smart
gadgets, each of which serves an important purpose in the home for the user and their
family. The IoT is the foundation of an intelligent home network, connecting disparate
intelligent devices such as smartphones, smart computers, and wearable devices. Citizens’
lives can be made easier and safer by making their homes more open and secure. The smart
home provides useful resources such as monitoring habits and even safety tests, which
have compelled consumers and system developers to conduct extensive research.
Blockchain-type systems and unified “cloud-like” computing networks can be used
to solve these problems. Blockchain was developed in 2008 by Satoshi Nakamoto and
includes a time-stamped set of malicious evidence documentation managed by a network
of autonomous networks [1]. Blockchain architecture consists of a series of blocks linked
together by simple cryptography. The three main concepts underlying the operation of
blockchain technologies are inflexibility, decentralization, and transparency. The three
roles have been remarkably effective, exposing them to a wide range of digital currency
technologies, such as the functionality of mobile vehicles, mobile phones, and embedded
systems. While the blockchain platform is secure and anonymous, there are some issues
https://www.mdpi.com/journal/sensors
Sensors 2022, 22, 4522 2 of 13

with its current implementation. For example, Sybil attacks by generations of false identities
to manipulate the community have become more complex.
Since standard methods only look at the signatures and do not operate on searching
for various specific patterns, a robust intrusion detection system is essential to analyze the
circumstances thoroughly. RTS-DELM is a machine learning technique used to analyze data.
This machine learning program uses an automated dataflow framework to determine data
flow to detect intrusions and attack patterns. To handle the continually emerging smart
blockchain-based applications, it is important to create powerful and versatile algorithms.
Machine learning is a method that includes computers that teach themselves using an
intelligent algorithm. According to one argument, machine learning is one of the first use
cases of Artificial Intelligence (AI). The theory of machine learning helps machines to solve
tasks without being explicitly programmed. The major objective of this sort of study is to
develop a realistic algorithm that will receive information from the input and forecast it, as
well as altering the outputs using statistical analysis. By utilizing machine learning, one
can process a massive amount of data and arrive at a judgment based on facts.
Furthermore, we suggest that a smart home network architecture that will overcome
the current problems related to the centralized security of home networks and will address
future attacks on smart home networks. In the present study, an RTS-DELM methodology is
used to make smarter homes safer using Internet of Things (IoT) powered sensors with en-
hanced efficiency. The key contributions to this research include a comprehensive overview
of technological innovations applicable to blockchain-based smart homes empowered by
RTS-DELM and a new outlook on diverse implementations (e.g., smart home data sharing),
assisted by the recent stages of technological advancement.
RTS-DELM enables the automation of data analytics procedures and the generation
of real-time insights. The datasets that the RTS-DELM methodology can assess can be
manipulated in smart home networks, which means that all inaccuracies can be eliminated.
Networks require consistent data. Any data-related issues in the RTS-DELM system will
be ignored. It contains a method for detecting and anticipating possible deception and
other unlawful activities. The purpose of this article is to examine an RTS-DELM-based
system model for the smart estimation of intrusion detection in smart home networks with
the highest degree of reliability. In the training and testing of intrusion detection in smart
home network optimization with RTS-DELM, the fused datasets NSL-KDD (123,323 data
samples) and KDD CUP 99 (25,193 data samples) are analyzed so that every instance
has specific and varied features. As a result, the analysis and comparison of the finest
approaches take place in the same place.
Due to the enormous amount of information transmitted in complicated networks,
data fusion algorithms can be of benefit. A data fusion operation can transform many
communications into useful and accurate data for the end-user. This article offers a data
fusion technique for networks that grow organically to huge numbers of nodes. Numerous
strategies and concepts for data fusion have been discussed in recent academic literature.
“Data fusion” and “information fusion” are two of the most often used categories. Our
investigation is limited to data supplied by sensors and not to data collected by any
other source. Sensors are employed in a data fusion method to improve the accuracy
of outcomes. The suggested approach secures the blockchain-based smart home system
by carefully evaluating its dependability concerning the important security objectives
of secrecy, authenticity, and usability. Furthermore, we assess our method’s ability to
efficiently secure sensitive information while using relatively little power to show our
argument that the overhead generated by our method is negligibly connected to the value
of its security and privacy.
The remainder of this article consists of the following parts. Section 2 contains survey
papers of related studies. Section 3 explains the underlying blockchain technologies and
introduces an RTS-DELM solution for blockchain-centered mobile home and smart home
application systems. Section 4 presents an approach to the RTS-DELM problem, includ-
Sensors 2022, 22, 4522 3 of 13

ing the simulation and findings of the DELM method. Finally, Section 5 addresses the
conclusions of the study found from the details.

2. Literature Review
Blockchain is a current buzzword among smart home enthusiasts, and a range of
research articles have been produced on the subject. In research exploring how blockchain
technologies would be used in the smart city, S. Aggarwal et al. [2] discussed many aspects
of healthcare, including transaction assimilation, home healthcare, and investment sharing.
Nonetheless, the application of blockchain technologies in the smart home has not been
explored comprehensively in information analysis. In many ways, the blockchain can be
used in the smart home sector. M. Andoni et al. [3] provided a detailed study of different
blockchain applications of a peer-to-peer resource sharing network. The study presents
detailed information on the implementation and capabilities of various smart home net-
works, such as security challenges in the smart grid, big data analysis, Artificial Intelligence
(AI), and payment systems. However, their study did not adequately account for smart
house-related issues, such as smart home security and smart city financial planning.
Khan et al. [4] suggested a user-based blockchain structure to secure the connectivity
of edge information in the Internet of Things. Z. Zhou et al. [5] researched blockchain tech-
nologies, contractual analysis, and distributed computing to transfer the control and per-
formance of certain automobiles. J. Wu et al. [6] suggested a software-specified blockchain
interface to recognize dynamic blockchain frameworks and proceeded to apply a consent
function approach to virtual machines with an application-aware system that can extract
and manage unique consensus resources.
Sivaraman et al. [7] examined security issues in the smart home networks and made
constructive suggestions. It is necessary to monitor and validate the systems that have been
approved with an algorithm server and run the smart home equipment also in the external
internet world. Via the abolition of a requirement for recognizing users and banning data
packets that are not originating from the internet, the current regulations cannot efficiently
protect internal user data. Lee et al. [8] introduced an upgrade that handles software
upgrades of embedded systems utilizing a blockchain, cryptographic certificates that are
implemented with shared keys, and introduces encryption protocols utilizing a private key.
A smart house has several tiny, embedded devices that are linked to each other. However,
to utilize this mechanism, there are more components required than what is needed for
a home.
Nevertheless, this method is less effective in the absence of adequate learning data to
generate a machine learning prototype. Panwar et al. [9] researched different cybersecurity
risks in the smart home and their responses. They offer a detailed description of smart
home defense in different forms of threats. Khan et al. [10] contributed a confidentiality
and integrity strategy. This sensor offers safe data collection, encryption, and inquiries for
smart home applications. It preserves the information exchanged between the individual,
gateway, network operator, and the system, thus promoting information verification and
privacy. Hsu et al. [11] proposed a multisensory data fusion technique within the smart
home methodology. They developed an intelligent smart home environment using smart
wearable devices to control and operate the services of the smart home.
In the current digital era, the exponential growth of Internet of Things (IoT) devices
raises various security and privacy-related design challenges for businesses. Previous
research indicates that blockchain technology is a key answer to the IoT’s data security prob-
lems. The blockchain method allows multiple data providers to share information safely
and reliably. IoT data are encrypted and stored in a distributed ledger [12]. Wang et al. [13]
investigated the security threats connected with data storage in sensor networks before
proposing the use of blockchain technology to guarantee the security of data stored in sen-
sor networks. The work provides both member proof and non-member proof by employing
a cryptographic accumulator instead of a Merkle hash tree. In addition, the number of
items in the current accumulator is restricted and cannot match the growing requirements
Sensors 2022, 22, 4522 4 of 13

of the blockchain. Therefore, the authors suggest a new form of the unbounded accumula-
tor, together with its description and security model. Finally, using bilinear pairings, an
unbounded accumulator system is constructed, and its performance is evaluated.
Bordel et al. [14] offer a theoretical framework for trust in IoT situations, including a
formalization in mathematics and a description of the conditions that a solution for trust
provision must satisfy. A study of these requirements reveals that blockchain technology
fully satisfies them; thus, a first trust provision system based on blockchain networks is
also presented. To evaluate the approach given, experimental validation is also proposed
and carried out. Ra et al. [15] suggested a method for password-protected secret sharing
key retrieval, secured by a secure password from a malicious key storage server of a PBC.
Existing systems for key recovery and password-protected secret exchange are described.
Singh et al. [16] proposed DeepBlockScheme: a deep learning-based blockchain-driven
scheme for a secure smart city, in which the blockchain is deployed at the fog layer to assure
manufacturing data integrity, decentralization, and security. Deep learning is used at the
cloud layer to boost productivity, automate data processing, and boost communication
bandwidth in smart factory and smart manufacturing applications. Zhang et al. [17]
presented a strategy that uses an expiry recognition approach based on the Least Recently
Used (LRU) algorithm to divide the blockchain transaction database into cold and hot
zones. It can save space by transferring unspent transaction outputs outside of in-memory
transaction databases. Salim et al. [18] proposed a quick, efficient handover authentication
technique that uses deep learning to authenticate devices and construct a user profile-based
system for instant permission. The channel state data of a user’s movement pattern train
the model and identify bad users impersonating honest users.

3. Proposed Methodology
A. Background
In the year 2008, Satoshi Nakamoto invented the blockchain. A blockchain-based
peer-to-peer payments system can eradicate intermediaries and double-spends by using a
primary cryptocurrency (e.g., bitcoins). It is a centralized method where any information
block is validated by SHA-256 (Secure Hash Algorithm) using a former hash block. The
block contains a large amount of transaction stored data such as the block number of the
previous block hash, transaction data, a nonce, and time stamps. The timestamp contains a
constant parameter, but a nonce parameter is randomly generated. The miner (computer
module) has control over the static (block) and dynamic (timestamp and nonce) portions
of an information chain and calculates the leading number of zeros required to form the
header of the block.
Figure 1 highlights the smart home network centered on the private blockchain net-
work. The platform includes four layers: a layer of data sources connecting to the network
and emitting user information, a private blockchain network layer empowered with RTS-
DELM that unleashes predictive analysis against thousands of files, a client node with
smart home system data, and a collection of home devices to simulate the data.
B. Integration Data Fusion Technique in Blockchain-Based Smart Home
Data fusion approaches incorporate information from various sensors to obtain more
accurate observations than could be accomplished by using a single, separate sensor.
Information extraction is the practice of extracting information from different and probably
interconnected sources and integrating it in a way that will get the most impactful outcomes.
For example, a series of network security sensors are in operation in a security framework.
It is impossible to achieve a wide-angle, all-encompassing image of the overall security
state of a dynamic system of the security system. Furthermore, devices that are spread
over an extensive range can be challenging to manage effectively. To increase the efficiency
of the model and provide analysis with an entire system protection condition, it is also
essential to effectively and intelligently fuse the outcomes of these sensors.
Sensors 2022, 22, x FOR PEER REVIEW 5 of 13
Sensors 2022, 22, 4522 5 of 13

Figure
Figure A proposedblockchain-based
1. proposed
1. A blockchain-based smart
smarthome
homenetwork.
network.
In comparison, several data sources can offer more consistent reliability because the
B. information
Integration Data
itself Fusion
comes fromTechnique in Blockchain-Based
various sources. Smart
Consequently, data Home
fusion approaches by
integrating information
Data fusion fromincorporate
approaches various datainformation
sources can from
produce more sensors
various reliable and stablemore
to obtain
results than those obtained by a single information source. In this manner, the NSL-
accurate observations than could be accomplished by using a single, separate sensor.
Information extraction is the practice of extracting information from different and
probably interconnected sources and integrating it in a way that will get the most
impactful outcomes. For example, a series of network security sensors are in operation in
a security framework. It is impossible to achieve a wide-angle, all-encompassing image of
Sensors 2022, 22, 4522 6 of 13

KDD [19] and KDD CUP 99 [20] datasets were used to evaluate the suggested system’s
performance—these datasets were utilized to perform data fusion. Each data collection
specifies a unique link that corresponds to a sequence of packets that flow between the
provider and target locations in the fused data collection according to a predefined protocol.
This data collection has 41 characteristics per record. Six distinct fields and 35 continuous
fields compose these features.
C. Integration of Real-Time Sequential Deep Extreme Learning Machine in Blockchain-
based Smart Home
Listed below are the advantages when RTS-DELM is implemented;
• User authentication, as a means to legitimately access and make transactions on the
blockchain network.
• Blockchain offers a high degree of trust and protection. Blockchain applications
incorporate real-time transaction mechanisms into smart contracts to ensure that the
contractual commitments, which were already negotiated, are fulfilled.
• Blockchain is a reliable way to incorporate an incentive-based mechanism to enable
consumers and users to make a data contribution. In addition, this big data would
help to refine the RTS-DELM model.
The usage of blockchain-based systems can be rendered smarter by the use of RTS-
DELM computational technology. The confidentiality of data can be enhanced when
utilizing the RTS-DELM distributed blockchain technology. RTS-DELM can also be used to
increase the pace at which comprehension is achieved by exchanging further knowledge,
thereby improving understanding. It offers the framework and network structure to
develop a decentralized blockchain application. In this article, we analyze the RTS-DELM
deployment architecture, which is an advanced system. The proper use of this technology
will be to collect intelligence from different information resources, such as sensors, mobile
devices, and IoT systems. Knowledge derived by utilizing these techniques is used for
smart apps. The blockchain is the central structural feature of smart apps. Nonetheless, for
analysis, the RTS-DELM method may be used to evaluate and forecast real-time data. The
blockchain also processes all the data that might be required from the RTS-DELM model.
Data errors such as duplication, missing data parameters, glitches, and noise are
reduced when making data for research. Knowledge is transmitted through the blockchain,
and the minimization of data-related issues can be solved in the RTS-DELM framework.
The RTS-DELM technique can function well where only a limited portion of a data set
is needed. The architecture provides a wide variety of implementations in various areas,
such as fraud detection and prevention. The blockchain infrastructure focuses on the
edge of the Internet of Things (IoT) and comprises three main elements: blockchain layer,
knowledge architecture, smart contracts, and the RTS-DELM framework. In the proposed
RTS-DELM system, vast quantities of hidden layers, hidden neurons, and several activating
mechanisms have been employed to optimize the privacy and protection of smart homes.
There are three separate steps in analyzing the data in the proposed method: the data
acquisition, preprocessing, and assessment stages. The evaluation layer was made up
of two sub-layers: the prediction and performance layer. For analysis, accurate data are
obtained from sensors and actuators. Then, the data are given as raw data and used
by the collection layer. A comprehensive technique for data cleaning and preparing has
been implemented to remove discrepancies in the preprocessing layer. The RTS-DELM
was employed to maximize the home network protection by preventing disruptive or
invasive applications.
Cryptographic hash functions connect blocks via cryptography. A home server com-
puter could be viewed as a miner to validate new transactions and introduce new blocks.
In contrast, intelligent contracts follow predefined rules to make decentralized transac-
tions easier and quicker. There are diverse consensus models for blockchains, such as
proprietary, public, and federated, but private blockchains are more effective when used in
smart homes.
Sensors 2022, 22, 4522 7 of 13

The deployment layer can provide interoperability between smart home devices and
blockchain networks. The paper tackles wireless home technology, interoperability in the
home, connectivity control, electronic billing of networks and municipal resources, and
healthcare. Consequently, the access layer serves to interact with devices at the intersection
layers such as at the individual level, the customer level (mainly for tech companies,
including for marketing and use in retail), as well as for the corporate level as the topmost
“layer” of the smart home, which primarily influences entities to take opportunity of the
blockchain smart home ecosystem, which is a huge market.
With smart appliances, smart technology, and smart upgrades, smart home architec-
ture is the tip of the iceberg; it is going to be a cornerstone of the smart home for smart
households. Living in a smart home environment includes having a smart scan of one’s
home with rooms, locking doors, smart devices that trigger conditions, and more. Through
the functions provided with the smart home, the user can regulate remote electricity, set
off alarms, watch and secure their home with video monitoring, and control their vehicle,
among several more applications. A user-friendly, custom-integrated framework that
would enable the homeowner to enable it on request should be introduced to increase the
seamless activity of smart homes and detect any disruptive behaviors by hackers. For an
IoT system, the user’s permission authorization can be listed in a collection of IoT system
control records, a specified set of IoT system owner records, or a specified set of IoT system
owner list records. Many sources that preserve trustworthy information must store those
pieces of information to withstand the assaults of hostile attackers.
To explain how blockchains will contribute to secure access, we present the following
explanation.
• First, the consumer must decide the access level and add it to the home service
computer. For example, at the highest level, the homeowner (Admin) is allowed, while
teenagers, youths, visiting relatives, and adolescents need mid-level permission.
• For a user who has access to the smart home and is using applications inside, Figure 2
shows how blockchain facilitates secure entry.
• Relatives and visitors have relatively poor access permits. When processing a request
from the client, the home server checks security access to repositories. Upon receiving
an order from the customer, the home server transmits the encrypted username and
password to the blockchain layer.
• For various users and implementations, a blockchain regulation header contains a set
of authorization rules. The part of the block data used for applying control policies
and services is the policy header.
The administrator checks the new user’s request and then accepts or rejects the access
request. When the information is incorporated into the blockchain, miners take action
depending on the policy specifics added to the header. This mechanism is successful in
combatting malicious attackers.
D. Real-Time Sequential Deep Extreme Learning Machine
RTS-DELM combines a variety of hidden layers, multiple hidden neurons, and a
variety of different activation functions to provide the ideal solution for improving smart
home networks. The proposed technique comprises three different phases: gathering the
data, reviewing it, and presenting it. We have two sub-layers in the application layer (there
are also additional layers in-between), where one is for estimation and the other layer is
for assessment. While experimenting, data were collected from sensors for observational
investigation. The data obtained through the data collecting mechanism were then made
available as input to the data collection mechanism. Until final processing, many data
processing techniques were used to eliminate anomalies from the results. Finally, the
RTS-DELM algorithm was utilized to improve smart home networks to prevent disruptive
or intrusive behavior.
EER REVIEW 8 of 13
Sensors 2022, 22, 4522 8 of 13

Figure 2. Proposed blockchain-based smart home management

Figure 2. Proposed blockchain-based system.system.
smart home management

The RTS-DELM method applies to a wide variety of smart home applications. To main-
D. Real-Time Sequential Deep Extreme
tain the requisite Learning
detection accuracy, Machine
a significant fraction of sensor readings is frequently
required. RTS-DELM addresses a variety of network access concerns through the use of
RTS-DELM combines a variety of hidden layers, multiple hidden neurons, and a
integrated routing and security measures. However, given that 80% of a network’s energy
variety of different activation
is used during functions to provide
data transmission the ideal
and reception, solution
data reductionfor
andimproving smart
function abstraction
home networks. Theapproaches
proposed may help to reduce
technique processing
comprises time different
three and increase the endurance
phases: gatheringof neural
the
networks. However, excessive compression will increase energy expenditures. Within
data, reviewing it, and presenting it. We have two sub-layers in the application layer
smart home networks, RTS-DELM enables more efficient data compression. As a result,
(there are also additional layers
intelligent in-between),
home networks where one
require real-time is forsolutions
networking estimation and scheduling,
for security, the other
monitoring, clustering nodes, data aggregation, and
layer is for assessment. While experimenting, data were collected from sensors forfault diagnostics. The RTS-DELM
design enables smart home networks to react smoothly to their changing environments.
observational investigation. The data obtained through the data collecting mechanism
In real-time, an RTS-DELM algorithm based on Deep Extreme Learning Machine
were then made available as input
(DELM) examines the datato the data
sequence. The DELMcollection mechanism.
may be used in a variety ofUntil final
applications
processing, many data and processing
domains to forecast health problems,
techniques were usedassess to
energy consumption,
eliminate inventory
anomalies services,
from the
and specify transportation activities. The RTS-DELM may be used to categorize and
results. Finally, the RTS-DELM algorithm was utilized to improve smart home networks
regressively dedicate data in a variety of ways since it is intuitive and efficient at keeping
to prevent disruptive or intrusive behavior.
The RTS-DELM method applies to a wide variety of smart home applications. To
maintain the requisite detection accuracy, a significant fraction of sensor readings is
frequently required. RTS-DELM addresses a variety of network access concerns through
Sensors 2022, 22, 4522 9 of 13

up with the intricacy of frameworks. An extreme learning computer is a feed-forward

neural network architecture that ensures that feedback can only shift one direction through
those various layers, but “back-propagation” through the initial neural network can be
consistently modified without error in the initial training process, where information passes
through the initial neural network and advances in reverse through the corresponding
error rate of achieving high precision at low errors. The weights in the model are constant
throughout the validation phase of the system, in which the validated model is imported,
the real data are predicted, and the precision of the model improves. There are several
hidden layers and at least one output layer in the RTS-DELM model, as well as an input
layer. Following conditioning, the framework is transported to the cloud for online use,
and it is then utilized for validation services throughout the cloud during the validation
process, as shown in the following diagram. In the assessment layer, the Mean Square Error
(MSE) was examined to enhance smart home networks and other connected devices.

4. Simulation Results
The Fused RTS-DELM method was applied to the fused dataset in this research. The
findings were distributed randomly between the training collection (126,238 samples)
and the test/validation set (15% of the tests) (22,278 records). The data were evaluated in
advance of their intended usage to ensure that they were error-free. The RTS-DELM method
needed to know if its machines had been affected by ransomware or other cyber threats.
Following that, we examined several neurons, including the activation of buried layers and
numerous types of active activities. In research designed to determine performance, we
successfully evaluated the production of RTS-DELM. To assess the RTS-DELM algorithm’s
performance, we used a variety of statistical measures that explained the result.
  
O1 O0 
V0 + V1
Misrate = (1)
V0 + V1
  
O0 + O1 
V0 V1
Accuracy = (2)
V0 + V1
O0 
V0
Sepecificity = (3)
O
( 0 V + O0 V )
 
0 1
O1 
V1
Sensitivity =    (4)
O1 + O1 
V0 V1
TP
True Positive Rate (TPR) = (5)
( TP + FN )
TN
True Negative Rate (TNR) = (6)
( TN + FP)
TP
Positive Prediction Value (PPV) = (7)
( TP + FP)
TN
Negative Prediction Value (NPV) = (8)
( FN + TN )
Table 1 shows the suggested RTS-DELM-based decentralized smart home network
framework to predict intrusion detection during the training level. Throughout the training,
a total of 126,238 recordings were used, divided into 65,495 and 60,743 normal and assault
records, respectively. As a consequence, it is seen that the forecasting system properly an-
ticipates 64,185 attack records of normal groups without a genuine attack while forecasting
1310 attack records incorrectly. In comparison, 60,743 records are obtained in the event of
Sensors 2022, 22, 4522 10 of 13

an attack, of which 58,707 are appropriately predicted as an attack and 2036 are incorrectly
forecasted as a regular activity while the attack occurs.

Table 1. Training of the proposed blockchain-based smart home network architecture for the estima-
tion of intrusion with fused dataset empowered with Fused RTS-DELM.

Suggested RTS-DELM-Based System Model

(85% of Data)
Total No. of Records (N = 126,238) Outcome (Output)
(Normal) (Attack)
Predictable outcome
O0 O1
V 0 = 65,495 64,185 1310
Input
Normal (TP) (FP)
V 1 = 60,743 2036 58,707
Attack (FN) (TN)

Table 2 shows the suggested RTS-DELM-based decentralized smart home network

framework to predict intrusion detection during validation level. Throughout the valida-
tion, a total of 22,278 records were used, divided into 11,558 normal and 10,720 assault
records. It is seen that 11,072 records of normal class with no attack identified are properly
predicted, whereas 486 records are forecasted falsely as having an attack, despite the ab-
sence of a genuine assault. In the event of a cyber-attack, 10,154 of the 10,720 data points
collected were properly predicted as invasions, whereas 566 were incorrectly predicted as
normal activities throughout the attack.

Table 2. Validation of the proposed blockchain-based smart home network architecture for the
estimation of intrusion with fused dataset empowered with Fused RTS-DELM.

Suggested RTS-DELM-Based System Model

(15% of Data)
Total No. of Records (N = 22,278) Outcome (Output)
Predictable outcome (Normal) (Attack)
(V 0 , V 1 ) O0 O1
V 0 = 11,558 11,072 486
Input Normal (TP) (FP)
V 1 = 10,720 566 10,154
Attack (FN) (TN)

Figure 3 shows the assessment of the proposed RTS-DELM-based decentralized smart

home network framework in terms of accuracy and the misclassification rate at the training
and validation levels. It was proven that the proposed RTS-DELM-based decentralized
smart home network system achieves a combined accuracy and misclassification rate
of 97.35% and 2.65%, respectively, during training. Additionally, during validation, the
proposed RTS-DELM decentralized smart home network system achieves a combined
accuracy and misclassification rate of 95.28% and 4.72%, respectively.
PEER REVIEW 11 of 13
Sensors 2022, 22, 4522 11 of 13

Figure 3. Different
Figure 3. Different statistical measures statistical measures
for the for the proposed
proposed blockchain-based
blockchain-based smart
smart home network
home network
architecture for the estimation of intrusion with fused dataset during validation and training.
architecture for the estimation of intrusion with fused dataset during validation and training.
Figure 3 illustrates the proposed blockchain-based smart home network security
system’sthe
Figure 3 illustrates effectiveness
proposed in various statistical parameters
blockchain-based smartduring the training
home networkand validation
security
phases using the Fused RTS-DELM proposed framework. This demonstrates unequivocally
system’s effectiveness in various statistical parameters during the training and validation
that the suggested method generates accuracy and misclassification rates of 97.35% and
phases using the2.65%, Fused RTS-DELM
respectively, proposed
during training. framework.
During validation, Thissystem
the suggested demonstrates
achieves an
unequivocally that the suggested method generates accuracy and misclassification rates
accuracy rate of 95.28% and a misclassification rate of 4.72%. Additionally, it demonstrates
the performance of the system model in terms of sensitivity, specificity, and true positive
of 97.35% and 2.65%, respectively, during training. During validation, the suggested
rate. (TPR), true negative rate (TNR), positive predicted value (PPV), and negative predicted
system achieves an accuracy
value (NPV) during ratetheof 95.28%
training and a phases.
and validation misclassification rate of 4.72%.
Additionally, it demonstrates the performance
We evaluated the dependabilityof of the system model
our technique in terms
on the reliability of sensitivity,
of other published
algorithms in the literature; in addition, as shown in Table 3, the suggested framework
specificity, and trueachieves
positive rate. (TPR), true negative rate (TNR), positive predicted value
much higher accuracy by reducing the error rate. In terms of accuracy, the
(PPV), and negativeproposed
predicted value framework
RTS-DELM (NPV) during the training
outperforms and validation
existing algorithms already in phases.
use, such
We evaluated the dependability of our technique on the reliability of other published
as the Artificial Neural Network-based Intrusion Detection System [21] and Generative
Adversarial Networks (GANs) [22] and Deep Extreme Learning Machine (DELM) [4].
algorithms in the literature; in addition, as shown in Table 3, the suggested framework
When compared to the DELM technique, the suggested RTS-DELM method achieves higher
achieves much higher accuracy
efficiency by dataset
on the fused reducing theof the
because error rate. accuracy
enhanced In terms of accuracy,
achieved the
[4]. In [21], the
proposed RTS-DELM framework
authors suggested an outperforms existing
Artificial Neural algorithms
Network-based already
Intrusion DetectioninSystem,
use, such
and, inas
this method, the researchers achieved 81.2% precision. In [22], the researchers suggested
the Artificial Neural Network-based Intrusion Detection System [21] and Generative
Generative Adversarial Networks (GANs), and in this method, the researchers achieved
Adversarial Networks 86.5%(GANs)
accuracy. [22] and
Finally, Deep
in the DELM Extreme
approachLearning
without data Machine (DELM)
fusion, 93.91% accuracy[4].
When compared to the DELM technique, the suggested RTS-DELM method achieves
higher efficiency on the fused dataset because of the enhanced accuracy achieved [4]. In
[21], the authors suggested an Artificial Neural Network-based Intrusion Detection
System, and, in this method, the researchers achieved 81.2% precision. In [22], the
Sensors 2022, 22, 4522 12 of 13

was achieved [4]. In this study, the RTS-DELM system achieved an accuracy of 95.28%,
which is greater than prior attempts, demonstrating its efficacy and demonstrating that the
system performance is improved by employing the data fusion technique. The suggested
RTS-DELM paradigm provides a much higher return on investment than existing methods.
As a result, the RTS-DELM paradigm that has been proposed gives a viable solution to the
aforementioned issue.

Table 3. Comparison of results of the proposed data fusion technique of decentralized smart home
network based on Fused RTS-DELM with the literature.

Method Accuracy Rate

ANN Based IDS [21] 81.2%
GAN [22] 86.5%
DELM [4] 93.91%
Fused RTS-DELM with Data Fusion (Proposed Blockchain Model) 95.28%

5. Conclusions
In this study, the idea of a smart contract in blockchain technology is employed to
validate the user’s identity for accessibility to centralized smart home services. The most
significant benefit of this research is the demonstration of how easy it is to receive facilities
and how secure the resources are. There is no need to have redundant authentication
because no other third-party users can access smart home systems, even if another user
tries to access an already used resource. Intrusion detection in smart homes, especially in
the context of assessment and prediction, remains a key concern. In the meantime, recent
advances in the blockchain and machine learning sectors have demonstrated tremendous
promise to accomplish these aims. Discussing the need for an efficient approach, this study
provided a compact and efficient mechanism for intrusion prevention. An RTS-DELM
approach was developed, and also data fusion techniques were presented to optimize
multi-sensor networks. Numerous measures were used to assess the feasibility of the
proposal. The consistency of RTS-DELM findings showed that the proposed method is
more successful than others. The suggested RTS-DELM solution obtained an exceptionally
high rate of success, showing 95.28% accuracy. The findings obtained are encouraging, and
we will continue to investigate more applications for the device through the deployment of
more datasets and varying frameworks.

Author Contributions: M.S.F. and S.K. collected data from different resources. M.S.F., S.K. and A.R.
performed formal analysis and simulation, M.S.F., S.K., A.R. and S.A. contributed to writing—original
draft preparation, M.A.K. and S.O.H.; writing—review and editing, M.A.K. and S.O.H.; performed
supervision, M.S.F., S.A. and A.R.; drafted pictures and tables, M.A.K. and S.A.; performed revision
and improved the quality of the draft. All authors have read and agreed to the published version of
the manuscript.
Funding: This work was supported in part by the Institute of Information and Communications
Technology Planning and Evaluation (IITP) under the High-Potential Individuals Global Training
Program under Grant 2021-0-01532 (50%), and in part by the National Research Foundation of Korea
(NRF) under Grant 2020R1A2B5B01002145 (50%), all funded by the Korean Government through
Ministry of Science and ICT (MSIT).
Institutional Review Board Statement: Not applicable for studies not involving humans or animals.
Informed Consent Statement: Not applicable for studies not involving humans.
Data Availability Statement: The simulation files/data used to support the findings of this study
are available from the corresponding author upon request.
Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2022, 22, 4522 13 of 13

References
1. Nakamoto, S. Bitcoin: A peer-to-peer electronic cash system. Decentralized Bus. Rev. 2008, 10, 21260.
2. Aggarwal, S.; Chaudhary, R.; Aujla, G.S.; Kumar, N.; Choo, K.K.R.; Zomaya, A.Y. Blockchain for smart communities: Applications,
challenges and opportunities. J. Netw. Comput. Appl. 2019, 144, 13–48. [CrossRef]
3. Andoni, M.; Robu, V.; Flynn, D.; Abram, S.; Geach, D.; Jenkins, D.; McCallum, P.; Peacock, A. Blockchain technology in the energy
sector: A systematic review of challenges and opportunities. Renew. Sustain. Energy Rev. 2019, 100, 143–174. [CrossRef]
4. Khan, M.A.; Abbas, S.; Rehman, A.; Saeed, Y.; Zeb, A.; Uddin, M.I.; Nasser, N.; Ali, A. A machine learning approach for
blockchain-based smart home networks security. IEEE Netw. 2020, 35, 223–229. [CrossRef]
5. Zhou, Z.; Wang, B.; Dong, M.; Ota, K. Secure and efficient vehicle-to-grid energy trading in cyber physical systems: Integration of
blockchain and edge computing. IEEE Trans. Syst. Man Cybern. Syst. 2019, 50, 43–57. [CrossRef]
6. Wu, J.; Dong, M.; Ota, K.; Li, J.; Yang, W. Application-aware consensus management for software-defined intelligent blockchain
in IoT. IEEE Netw. 2020, 34, 69–75. [CrossRef]
7. Sivaraman, V.; Gharakheili, H.H.; Vishwanath, A.; Boreli, R.; Mehani, O. Network-level security and privacy control for smart-
home IoT devices. In Proceedings of the 2015 IEEE 11th International Conference on Wireless and Mobile Computing, Networking
and Communications (WiMob), Abu Dhabi, United Arab Emirates, 19–21 October 2015; pp. 163–167.
8. Lee, B.; Malik, S.; Wi, S.; Lee, J.H. Firmware verification of embedded devices based on a blockchain. In International Conference on
Heterogeneous Networking for Quality, Reliability, Security and Robustness; Springer: Cham, Switzerland, 2016; pp. 52–61.
9. Panwar, N.; Sharma, S.; Mehrotra, S.; Krzywiecki, Ł.; Venkatasubramanian, N. Smart home survey on security and privacy. arXiv
2019, arXiv:1904.05476.
10. Khan, M.A.; Ghazal, T.M.; Lee, S.W.; Rehman, A. Data Fusion-based machine learning architecture for intrusion detection. Comput.
Mater. Contin. 2022, 70, 3399–3413.
11. Hsu, Y.L.; Chou, P.H.; Chang, H.C.; Lin, S.L.; Yang, S.C.; Su, H.Y.; Chang, C.C.; Cheng, Y.S.; Kuo, Y.C. Design and implementation
of a smart home system using multisensor data fusion technology. Sensors 2017, 17, 1631. [CrossRef] [PubMed]
12. Le Nguyen, B.; Lydia, E.L.; Elhoseny, M.; Pustokhina, I.; Pustokhin, D.A.; Selim, M.M.; Nguyen, G.N.; Shankar, K. Privacy
preserving blockchain technique to achieve secure and reliable sharing of IoT data. Comput. Mater. Contin. 2020, 65, 87–107.
[CrossRef]
13. Wang, J.; Chen, W.; Wang, L.; Sherratt, R.S.; Alfarraj, O.; Tolba, A. Data secure storage mechanism of sensor networks based on
blockchain. Comput. Mater. Contin. 2020, 65, 2365–2384. [CrossRef]
14. Bordel, B.; Alcarria, R.; Martin, D.; Sanchez-Picot, A. Trust provision in the internet of things using transversal blockchain
networks. Intell. Autom. Soft Comput. 2019, 25, 155–170. [CrossRef]
15. Ra, G.J.; Roh, C.H.; Lee, I.Y. A key recovery system based on password-protected secret sharing in a permissioned blockchain.
Comput. Mater. Contin. 2020, 65, 153–170. [CrossRef]
16. Singh, S.K.; Azzaoui, A.E.; Kim, T.W.; Pan, Y.; Park, J.H. DeepBlockScheme: A deep learning-based blockchain driven scheme for
secure smart city. Hum. Cent. Comput. Inf. Sci. 2021, 11, 12.
17. Zhang, J.; Zhong, S.; Wang, J.; Yu, X.; Alfarraj, O. A storage optimization scheme for blockchain transaction databases. Comput.
Syst. Sci. Eng. 2021, 36, 521–535. [CrossRef]
18. Salim, M.M.; Shanmuganathan, V.; Loia, V.; Park, J.H. Deep learning enabled secure IoT handover authentication for blockchain
networks. Hum. Cent. Comput. Inf. Sci. 2021, 11, 21.
19. Dhanabal, L.; Shantharajah, S.P. A study on NSL-KDD dataset for intrusion detection system based on classification algorithms.
Int. J. Adv. Res. Comput. Commun. Eng. 2015, 4, 446–452.
20. Tavallaee, M.; Bagheri, E.; Lu, W.; Ghorbani, A.A. A detailed analysis of the KDD CUP 99 data set. In Proceedings of the 2009
IEEE Symposium on Computational Intelligence for Security and Defense Applications, Ottawa, ON, Canada, 8–10 July 2009;
pp. 1–6.
21. Ingre, B.; Yadav, A. Performance analysis of NSL-KDD dataset using ANN. In Proceedings of the 2015 International Conference
on Signal Processing and Communication Engineering Systems, Guntur, India, 2–3 January 2015; pp. 92–96.
22. Alshinina, R.; Elleithy, K. A highly accurate machine learning approach for developing wireless sensor network middleware. In
Proceedings of the 2018 Wireless Telecommunications Symposium (WTS), Phoenix, AZ, USA, 17–20 April 2018; pp. 1–7.